Deficiency Taxonomy - Categorizing Deficiencies within my Organization

Mnts2C

Starting to get Involved
#1
Hello everyone, I’m SO glad to see the Cove back !!! Even though I haven’t met anyone in real life, just seeing the familiar experts made me very happy to see their personalities again and of course, the site is back.

My challenge lies in categorizing deficiencies within my organization. We have a software program where we enter information regarding deficiencies caused by us. The data is mainly of who, what, when, where, etc. When the user gets to actually ‘classifying’ the deficiency the method involves completing roughly nine blocks or buckets. These nine blocks consist of three ‘areas’ (Metric, Process, & Program) and three levels of detail (~ gross, medium, fine). Our guide to help classify is attached and a crude example is below. Note that the guide gives no procedure for HOW to classify a deficiency, only the available buckets.

Deficiency example = Use of non-calibrated torque wrench during joint reassembly.

Metric Process Program
  • 1st Lvl WS-Workmanship 08-Production S&EP-Steam & Electric Plant
  • 2nd Lvl WSb-Skill/Technique 08c-Mechanical None given
  • 3rd Lvl WSb08-Assembly Incorrect 08c04-Mech Jt Make-up None given

In case the above reads poorly, I've attached my question as a Word file.

So this is a pretty generic example and often enough we have deficiencies that cause us to scratch our heads with all the options available to classify. Using the same deficiency, many folks may classify the deficiency differently. Thus causing variation in the data and thereby negatively affecting any type of analysis.

My thought to minimize this variation is to create a structured methodology for deficiency classification. The HOW. Call it a taxonomy. Scientist discover a new animal and they use a rigorous method that gives us the Kingdom, Phylum, Class, Order, Family, Genus, Species to classify that new bug or whatever.

For our deficiencies, we do not have that methodology. Would I start with Process first if the deficiency was discovered during test, then move to Program? The Metric areas roughly mimic the “when” the deficiency occurs during a normal job flowpath (Tech Direction (paper), Work Readiness (parts), Work Controls (power isolated), Workman ship (performing to the paper), Testing (prove the job), and Certification (verify acceptability of the whole job)). Should I start with Metric then go to Program?

I’m looking for some guidance to create some type of process flowchart that would improve classification accuracy.

Thoughts ??

Thanks in advance !!
 

Attachments

Elsmar Forum Sponsor

Mike S.

Happy to be Alive
Trusted Information Resource
#2
Just off the top of my head, it looks to me like you are trying to get too much detail too early in the data analysis process. If you try to get this much detail on every deficiency, you will create a lot of work and confusion before you even know if the class of deficiency needs attention.

Here's my take: Think Pareto. Think simple. Think large buckets for deficiencies like wrong tool used, inadequate process/procedure, missing hardware, FOD, finish problem, etc.

Versus narrow buckets like wrong torque wrench, wrong screwdriver, uncalibrated torque wrench, uncalibrated voltmeter, uncalibrated oscilloscope, etc.

Find the large bucket with the most problems, then as needed narrow the problems down to the point you can address the problems as appropriate.
 

Mnts2C

Starting to get Involved
#3
Thanks Mike. I understand the Less is More philosophy you are describing and believe in its value when used appropriately, truly I do. And I have to work with the tools available which include our deficiency data program where we receive ~60,000 entries per year (I have to play the hand I’m dealt). So without going into other problems with the program, I’m trying to focus on creating a classification methodology for deficiencies.

Even removing the lowest level of detail, leaving only the high and medium levels of classification, I still remain stumped for developing a methodology. Let me throw out something for consideration.

Scenario is a large manufacturing and repair facility of ~14,000 people. Items are made that support different programs; items for underwater use, mechanical items, nuclear items, electrical items, items for outer space applications, etc. And all the associated stuff; logistics, parts procurement, administrative processes, receipt inspection, developing engineering work documents, variety of testing functions, metrology, project management, design services, training, oversight, etc.

Here’s a potential new deficiency: An o-ring of incorrect material/durometer is installed in a complex assembly and discovered after the assembly is in use by the customer. Using the previously attached definition guide to classify the deficiency and the deficiency scenario above, I want to create a type of flowchart that allows multiple users to arrive at a common classification of the deficiency.

Should the classifier start with the Program category?

Step 1: Identify the Program that the deficiency falls under. (We can only enter one Program in the database. What if the assembly is used in underwater and outer space programs, what is the decision point? Notice the guide has a Material Control program as well; someone could choose Material Control because an incorrect o-ring was installed. Or if the receipt inspection was faulty, the Receipt Inspection program could be chosen.)

Step 2: Identify the sub-Program category that the deficiency falls under, as applicable.

Step 3: Identify the Process that the deficiency falls under. (If engineering called out the incorrect o-ring and receipt inspection failed to identify the error, and the worker failed to verify the o-ring to the drawing, which Process is selected? Engineering, Inspection, or Production?)

Step 4: Identify the sub-Process category that the deficiency falls under, as applicable.

Step 5: Identify the Metric that the deficiency falls under. The Metric category essentially equates When along the job timeline. (If engineering called out the incorrect o-ring and receipt inspection failed to identify the error, and the worker failed to verify the o-ring to the drawing, which Metric is selected? Technical Direction, Work Readiness, or Workmanship?)

Step 6: Identify the sub-Metric category that the deficiency falls under, as applicable.
I’m aware our deficiency categories in our definition guide need improvement and we are working on that. Even with useful and non-redundant categories, the method of HOW someone selects them is vital. I believe without the method flowchart, ten different classifiers would not arrive at the deficiency categories.

The inability to classify deficiencies consistently results in poor/invalid/inaccurate data. This affects evidenced based decision making when the evidence is questionable. Should management improve the Receipt Inspection program or should they improve the Training process of engineers, or improve the o-ring design process?

A long post and I thank those patient enough to read and reply.
 
R

randomname

#4
Google "orthogonal defect classification." It's related to software bugs but might give you some ideas.
 

Similar threads

Top Bottom